AI lung cancer predictor finds patients most likely to relapse

Ross Lydall

London researchers are using artificial intelligence to identify the lung cancer patients at higher risk of having the disease return after treatment.

Lung cancer is the most common cancer death in the UK, killing more than 35,000 people a year.

Though in its early stages, the breakthrough could help doctors to offer close monitoring and tailored treatment to improve survival rates.

The work, part of Cancer Research UK’s landmark £14 million, nine-year TRACERx study, has been carried out at the Institute of Cancer Research, University College London and the Francis Crick Institute.

Computer scientists and pathologists created a AI system to differentiate between immune cells and cancer cells, thus showing how lung tumours evolve in different patients.

Areas packed with immune cells were described as “hot” and those devoid of the cells were defined as “cold”. Patients whose tumour had more “cold” regions were more likely to relapse, according to analysis of samples from 100 patients published in Nature Medicine.

The team, led by Dr Yinyin Yuan from the ICR, found that cancer cells in cold regions might have evolved more recently than cancer cells found in hot regions.

This suggested that the tumour might have developed a “cloaking” mechanism to hide from the body’s natural defences.

Dr Yuan said: “Cancer’s ability to evolve and to come back after treatment is one of the biggest challenges facing cancer researchers and doctors today.

“Our research has revealed fresh insights into why some lung cancers are so difficult to treat.”